From unbuilt
Use when you or the user deliberately decide NOT to build something, cut a feature to keep things simple, rule out an approach, drop a dependency, or set a scope boundary — phrases like "we're not doing X", "leave that out", "keep it simple here", "that's out of scope", or "YAGNI". Records the decision and its reasoning in the project's UNBUILT.md so it is not silently reversed or re-litigated later. Trigger this right after such a decision is made.
How this skill is triggered — by the user, by Claude, or both
Slash command
/unbuilt:record [what you're choosing not to build, and why][what you're choosing not to build, and why]The summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are adding one entry to the project's deliberate-omissions ledger,
You are adding one entry to the project's deliberate-omissions ledger,
UNBUILT.md, at the root of the current project. This is the record of things
the project chose not to build, and why — the mirror image of a decision
record.
Work out the entry.
$ARGUMENTSRead the existing ledger. Look for UNBUILT.md in the project root.
Append the new entry — never overwrite.
YYYY-MM-DD).Entry format:
## <short title> — <YYYY-MM-DD>
**Left unbuilt:** <what is deliberately not being built, or the boundary>
**Why:** <one-line reason>
**Revisit if:** <condition that would reopen this, or "—">
If you are creating the file, start it with this header, then the first entry:
# Unbuilt
Things this project deliberately does **not** do, and why — so they are not
silently re-added later. Newest entries at the bottom.
Confirm. Tell the user, in one line, what you recorded. If the write fails (for example, a read-only folder), say clearly that the entry was not saved.
UNBUILT.md at the project root (the top of the current
working tree). Touch nothing else.Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
npx claudepluginhub gru-953/unbuilt --plugin unbuilt